Primer dimer contamination

[buthercup_ch]

Hello,

I would like to ask for the opinion of those having experience in this procedure.
We are dealing with bacterial transcriptomics and applying RNA-Seq.

We are following the TruSeq RNA Sample preparation guidelines from Illumina, the Low Sample (LS) protocol, and starting from 0.5 ug of total RNA samples.
Ribo-Zero Magnetic Gold kit was used as starting point to deplete samples from ribosomal RNA.

Ok, so the first issue:
1) the quantification by Qubit of the final libraries gave very low values, ranging form 1.22-7.3

and the second:
2) there was a huge contamination of what seems to be primer-dimer (see attachment)

I am struggling trying to understand where these problems might come and how to solve them in the next preparation, as it is clear that the present libraries cannot go into the sequencer.

Thanks in advance for your comments

Cristina

[jdk787]

The peak at 120bp is adapter dimer which could be caused by overestimating your starting material concentration, using a higher concentration of adapters than recommended, or incorrect bead volumes for the cleanups.

Other than that these libraries look ok, so I would try performing an additional 0.8X bead cleanup on hour final libraries to remove the adapter dimers.

[buthercup_ch]

Hi jdk787,

Thanks for your answer.
Do you mean to use AMPure XP beads once more?
Just to clarify, 0.8X means the ratio of volume to be mixed right? 0.8 vol beads + 1 vol library?

I am also worried about the low yield, as I am used to obtain about 30-50 ng/ul.

Is not true that I might loosing material, let's say those low abundant transcripts, and the results will be hence biased?

Thanks

[jdk787]

Hi jdk787,

Thanks for your answer.
Do you mean to use AMPure XP beads once more?
Just to clarify, 0.8X means the ratio of volume to be mixed right? 0.8 vol beads + 1 vol library?

I am also worried about the low yield, as I am used to obtain about 30-50 ng/ul.

Is not true that I might loosing material, let's say those low abundant transcripts, and the results will be hence biased?

Thanks

0.8X will be 0.8 vol of bead to 1 vol of library. So for a 50uL sample you would add 40uL of beads

30-50 ng/uL is a high yield for a final library (over 100nm for your size)
A 2ng/uL library that is in the size range that you have will be ~10nm, which is sufficient for sequencing.

I am curious about what the BA says your concentration is because based on the amount of adapter dimer alone I would think that your samples should be greater than 2ng/uL.

[nucacidhunter]

Library profiles shows characteristics of low input library prepared with a kit designed for high input. If input RNA quality was good then any failure or low efficiency during library prep steps would cause this including issues with library prep kit. I would be hesitant to sequence those if the aim is differential expression analysis.

[buthercup_ch]

0.8X will be 0.8 vol of bead to 1 vol of library. So for a 50uL sample you would add 40uL of beads

30-50 ng/uL is a high yield for a final library (over 100nm for your size)
A 2ng/uL library that is in the size range that you have will be ~10nm, which is sufficient for sequencing.

I am curious about what the BA says your concentration is because based on the amount of adapter dimer alone I would think that your samples should be greater than 2ng/uL.

Thank you so much jdk787 for your clarifications.

We will try to clean-up the libraries once more. I hope we'll not loose too much material in the process.

[buthercup_ch]

Library profiles shows characteristics of low input library prepared with a kit designed for high input. If input RNA quality was good then any failure or low efficiency during library prep steps would cause this including issues with library prep kit. I would be hesitant to sequence those if the aim is differential expression analysis.

hello nucacidhunter,
thanks for your comment.

Please, I am getting a bit lost with the vocabulary... what do you mean by "low input" and 'high input"?

Total RNA samples were of high quality upon verification on Bioanalyzer and Nanodrop.

I really thing the problem was caused by a non efficient clean-up of the adaptor, so it got highly amplify during PCR enrichment.... That's is my hypothesis.

[GenoMax]

Please, I am getting a bit lost with the vocabulary... what do you mean by "low input" and 'high input"?

Low concentration of input sample. With a kit that is designed for/expects a more concentrated sample.

[nucacidhunter]

Your hypothesis is probable. I do not know how well the protocol has been followed. Non-smooth curves of samples 2and3 and massive primer dimers are similar to libraries with lower than required input quantity (for instace using 10 ng input when the kit calls for 100 ng). Some other samples profiles have been dwarfed by large dimer peaks so they could also have rough surface. With Truseq RNA kits yields above 10 ng/ul are common and low yields are good indicators that something has been non-optimal.

[buthercup_ch]

Thanks nucacidhunter for the clarification.

We started with 500 ng... so I guess a low input was not the case of "failure"... I did not perform the preparation, but I guess a combination of low cleaning-up of the adapter and a low number or cycles during PCR enrichment could cause the low yield of the final library (<10 ng/ul).

We are going to clean-up again and repeat the quantification (that might also be not very accurate). I hope we'll not loos too much.

In any case, if final yield is <<10 ng/ul... in your opinion, you will not proceed with sequencing?
Thanks

[nucacidhunter]

In my experience those dimer peaks will not be totally eliminated even with two clean up and will comprise high proportion of reads. I think a bioinformatician should be consulted to see if they can normalise library prep variation in addition to standard analysis. These libraries do not look similar and they will contribute to observed expression differences in addition to treatments or tissue type. Yield is one of indicators of library quality.